The capacity of neural stem cells (NSC) to migrate toward pathological areas of the brain underscores the potential use of these cells as agents for cell replacement and/or drug delivery in the brain. Malignant gliomas consist of cancer stem cells (CSC), which have been demonstrated recently as the roots of the tumor, being refractory to current employed therapies. We have described the efficacy of using primary murine fetal NSC or bone marrow-derived neural stem cells (BM-NSC) as delivery vehicles for cytotoxic or immunostimulatory agents to treat infiltrating glioma and have described a mechanism of glioma tropism. However, these studies were based on tumor models established with glioma cell lines. Increasing evidence indicates that primary tumor cells are necessary and better than tumor cell lines for studying tumor biology and therapeutic strategies. We have recently demonstrated that primary human glioblastomas contain CSCs. The CSCs have characteristics of self-renewal, are multipotential in vitro, and can initiate brain tumors in vivo, recapitulating the phenotypes of the primary tumors from which they were derived. We also found that BM-NSCs can migrate toward CSCs in vitro and CSCs initiated tumors in vivo. These findings lead us to pursue further studies with adult BM-NSCs as a viable source of cellular vectors to target brain tumor CSCs and deliver therapy. We now aim to test the hypotheses that: 1) BM-NSC migration toward CSCs of primary brain tumors is dependent on CXCR4 expression on the plasma membrane of BM-NSCs. 2) BM-NSC migration toward brain tumor CSCs is mainly due to GFAP+ and A2B5+ astrocytic precursors, while terminal differentiation favors engraftment of BM-NSC after intracranial transplantation. 3) Targeting CSCs with BM-NSCs can (i) translate into targeted delivery of therapeutic genes and (ii) increase tumor control and prolong survival in a CSC xenograft model. Brain tumors are among the most devastating tumors and are often rapidly fatal despite aggressive treatments. Brain tumor cancer stem cells identified from human brain tumors have the exclusive ability to drive tumor formation, and could prove an effective target for tumor therapy. In this proposal, we will develop therapeutic strategies that are effectively targeting cancer stem cells and will ultimately yield new approaches to treat brain tumors.